Roadmap

What exists. What's next.

Biomata Engine and the Unity SDK are open source and available today. Everything else on this page is planned — not available. We're honest about that.

Read the Docs View on GitHub

Available Now · Open Source · Apache 2.0

Runtime

Biomata Engine

Execution infrastructure for autonomous agents.

The core Python simulation runtime. Manages cognition loops, memory systems, world abstraction, event orchestration, and simulation state. The engine beneath every simulation.

  • Agent cognition loops with configurable tick rates
  • Pluggable Brain protocol — LLM, rule-based, scripted, or RL
  • Memory systems (working, episodic)
  • World abstraction layer — local or host-authoritative
  • Deterministic event bus and scheduler
  • Simulation state snapshots and replay
  • WebSocket Protocol v1 server and client
  • YAML config — import any class via dotted path
Documentation → Source on GitHub →

SDK

Unity SDK

C# package for Unity 6. Host-authoritative integration.

Drop the package into a Unity 6 project. BiomataManager connects to the Python runtime, registers agents, drives the tick loop from FixedUpdate, and applies engine commands to GameObjects.

  • BiomataManager MonoBehaviour — connection lifecycle
  • AgentClient — per-agent observation sending and command receipt
  • TickClient — synchronize cognition to Unity FixedUpdate
  • WebSocket Protocol v1 implementation in C#
  • Event system — OnDecision, OnAgentRegistered, OnDisconnect
  • Engine Owned and Host Owned sample scenes
Unity SDK docs → Source on GitHub →

WebSocket Protocol v1 — open spec, any engine

The protocol is engine-agnostic and fully documented. Unity implements it. You can implement it for Godot, Bevy, Unreal, or a browser client without waiting for an official SDK.

Protocol Spec →

Planned — Not Yet Available

Tooling layer

Visual design, behavioral observability, and deterministic benchmarking. Planned for a future release.

Biomata Studio

Planned

Design and inspect living simulation systems.

A visual environment for composing worlds, placing agents, wiring behaviors, and inspecting systems at runtime.

Biomata Observe

Planned

Understand why agents behave the way they do.

Deep behavioral observability. Trace decision chains, visualize cognitive state, and surface causal paths behind agent actions.

Biomata Bench

Planned

Measure autonomous behavior rigorously.

Deterministic evaluation for agent decision quality. Define metrics, run reproducible benchmarks, track regression over time.

Planned — Not Yet Available

Scale & Ecosystem

Distributed execution, experiment orchestration, and a community ecosystem. Long-term direction — not accepting contributions yet.

Biomata Cloud

Planned

Distributed simulation execution.

Run simulations at production scale across distributed infrastructure without changing your simulation code.

Biomata Orchestrator

Planned

Large-scale experiment coordination.

Coordinate parallel experiment runs: define parameter sweeps, manage run state, collect and aggregate results.

Start with what's here.

The engine and Unity SDK are open source and ready to use today. Everything else is planned.

Read the Docs GitHub